Appendix S1.Method for Combining the Bird Point-Count and Mist Net Datasets

Appendix S1.Method for combining the bird point-count and mist net datasets

Capture and observation methods may yield differing results in terms of bird responses to disturbance (Hill & Hamer 2004; Barlow et al. 2012). For example, point counts are generally more efficient in mature forest, whereas mist nets are more efficient in younger forest (Hill & Hamer 2004). We therefore combined the two datasets. To do this, where species were recorded by both methods, we assigned them to just one for analysis, based on family level taxonomy. Species that were recorded by just one method were included in the analysis irrespective of taxonomy. Based on species accounts in Restall, Rodner & Lentino (2006), all families that dwell primarily in the canopy (i.e. mid, upper and above canopy; parrots, woodpeckers, cotingas, oropendulas, toucans, tanagers, trogons, puffbirds, grosbeaks, greenlets, jacamars, motmots) were assigned to point counts, as well as understorey and ground dwelling birds with body sizes too large to be reliably sampled by mist nets designed for smaller birds (currasows and guans, trumpeters, tinamous, doves and pigeons,cuckoos, raptors). Understorey dwelling families were assigned to mist nets (woodcreepers, ovenbirds, manakins, wrens, hummingbirds, thrushes, kingfishers), with the exception of antthrushs and antpittas which were assigned to point counts as they were infrequently captured in mist nets, but regularly recorded by point counts. Two diverse families; typical antbirds (Thamnophilidae) and flycatchers (Tyrannidae), included both upper-level and understorey dwelling birds. As point counts sample a greater range of species, with higher overall detection rates, we assigned these groups to point counts, with the exception of one genus of flycatchers (Mionectes spp.) which were assigned to mist nets as they were readily caught in mist nets, but not often recorded by point counts.
Table S1.Predictor variables at the 17 sample sites subject to Reduced-Impact Logging, and three primary forest control sites. Time elapsed between pre- and post-logging surveys in days, time elapsed between logging and post-harvest surveys in months, logging intensity (trees and volume extracted), and skid trail density. Logging variables were measured within a 200m radius from the centre of the sample site. Mean density of trees above 40 cm diameter before logging was 403 trees ha-1.

1

Table S2. Significant (P< 0.05) bird and bat indicator species (IndVal:Dufrêne & Legendre 1997), and their association with the NMDS ordination axes. Difference shows the disparity between the rtaucorrelation with Axis 1 and Axis 2, whereby species with the greatest differences exhibited responses to either background rates of turnover (+ve) or Reduced-Impact Logging (-ve); species in bold are those mentioned in the text. IV rates the value of the species as an indicator. Abundance post-logging shows whether the species was more abundant pre- (-) or post-logging (+).

aFr=Frugivore; In=Insectivore; Om=Omnivore; Ca=Carnivore; Un=Understorey; Up=Upper-levels

1

Figure S1.Bird (A) and bat (B)individual based rarefaction curves pre- and post-Reduced-Impact Logging for all species, and separately for frugivores and insectivores, conducted in EstimateS 8.2.0 (Colwell et al. 2006). Shaded areas indicate 95% confidence intervals of the observed species richness. Darker colours are pre-logging, lighter colours post-logging. The curves end at the total number of individuals pre-logging.The dot represents the total number of captures post-logging, followed by the extrapolated richness (dashed line) using the method in Colwell et al. (2012).

1

Figure S2. Change in abundancefor each species of bird (A), bat (B) and large mammal (C) recorded in the studypre- and post-Reduced-Impact Logging . Grey bars show the pre-logging abundance, green bars show increase in abundance post-logging, and red bars show decrease in abundance post-logging, in which case the cumulative red and grey represent the pre-logging abundance. Labelled species are those identified as indicators (Table 3) by IndVal. Species in bold are those mentioned in the text. Species are ordered within their family and guild according to their abundance pre-logging. Families represented by 4 species, are placed in the various ‘other’ categories.

1

Supporting references

Barlow, J., Silveira, J.M., Mestre, L.A.M., Andrade, R.B., D'Andrea, G.C., Louzada, J., Vaz-de-Mello, F.Z., Numata, I., Lacau, S. & Cochrane, M.A. (2012) Wildfires in Bamboo-Dominated Amazonian Forest: Impacts on Above-Ground Biomass and Biodiversity. Plos One,7.

Colwell, R.K. (2006) EstimateS: Statistical estimation of species richness and shared species from samples. Version 8.2.0.

Colwell, R.K., Chao, A., Gotelli, N.J., Lin, S.Y., Mao, C.X., Chazdon, R.L. & Longino, J.T. (2012) Models and estimators linking individual-based and sample-based rarefaction, extrapolation and comparison of assemblages. Journal of Plant Ecology,5, 3-21.

Dufrêne, M. & Legendre, P. (1997) Species assemblages and indicator species: The need for a flexible asymmetrical approach. Ecological Monographs, 67, 345-366.

Hill, J.K. & Hamer, K.C. (2004) Determining impacts of habitat modification on diversity of tropical forest fauna: the importance of spatial scale. Journal of Applied Ecology,41, 744-754.

Restell, R., Rodner, C. & Lentino (2006) Birds of Northern South America: An identification Guide. Christopher Helm Publishers Ltd.

1